The Question of Randomness

[TheNegotiator 13]

When I first came across this ages ago it made me sit up and it did the same when I came across it again recently. Take a look at the chart I have attached. What is it you ask? Oil, gold, S&P 500. Nope, it's nothing. It's a chart which I had excel generate simply by getting it to choose a tick higher, a tick lower or unchanged from the last price. But it looks very familiar doesn't it?

 

Now we all know the markets aren't random , but really when you look at an example like this it makes you wonder. What are the implications of such an exercise?

 

Here are some thoughts:

 

1. There is an element of perceived randomness to the market at times which may or may not perpetuate further 'random' activity.
    
   
2. Markets are not random at all as conditions present at any given moment, caused the subsequent price movement. However, when you look at historical charts, they have the appearance of being randomly generated as there is no context when viewing historical prices.
    
   
3. Unless we assign useful context to market movements, any analysis which is done is never going to be much better than random.
    
   
4. Would you trade the product in my chart if you knew it were completely random???
   

 

What do you guys reckon?

[TheNegotiator 13]

Here's another. Btw, just to be clear - I didn't cherry pick either chart.

[Do Or Die 10]

What primarily distinguishes random graphs to real price history is outliers- in reality, markets often stagnate at one level or may move in one direction for extended period of time.

 

Plotting an ROC for a random chart vs real chart will show remarked difference.

 

Also, if you plot candlesticks charts with the random data, any trader using candlestick charts will be able to tell the difference in a wink

[TheNegotiator 13]

I'm not so sure about whether it'd be easy to tell the difference. I'd like to do a test of this but I'm afraid I don't know how to easily group my random tick data into large groups for OHLC candles or bars in my Excel spreadsheet.

[TheNegotiator 13]

Okay so here are three(non-cherry picked) 20 bar samples of a 144 tick chart of the random data.

 

Just remember this is a very simple exercise. Each 'tick' which is randomly generated here can give rise to a change in price which is clearly not going to be as likely in a real product. But it does give a nice illustration.

Edited September 6, 2011 by TheNegotiator

[Do Or Die 10]

When I used to trade on floor, a friend once came with several candlestick charts which were generated out of random data.

 

They were intimidating at first, but we all came to the conclusion that none of the classic chart patterns can be seen on random charts. For example, a extended up run which culminates in 'reversal bar', 'long-legged doji' or 'engulfing bear'.

 

BTW, could you please post a chart with minimum 100 bars... 20 bars is too short of 'price' history.

Edited September 6, 2011 by Do Or Die
minimum 100 bars

[TheNegotiator 13]

To start with, I don't know how this data was generated. It certainly makes a difference. Secondly, you were obviously intrigued enough to look further into it, which then gives rise to working out if there are any important ideas to draw from it. I have said already I don't believe markets are random.

 

Yes I know 20 bars is too short and no I won't post one with more than 100 unless someone can do it in excel for us as I did the exercise manually!!

[Do Or Die 10]

I'm not sure which formula you used to calculate the candlesticks... but if you calculate each of O/H/L/C totally independent from any previous data, the candlesticks get pretty messed up.

[TheNegotiator 13]

I used a starting price to calculate each sequential tick - either 1 higher, 1 lower or unchanged. Then I grouped the 'ticks' into 144 and took the open, the close, the max and the min values from each range.

[Tradewinds 12]

 

1. There is an element of perceived randomness to the market at times which may or may not perpetuate further 'random' activity.
    
   
2. Markets are not random at all as conditions present at any given moment, caused the subsequent price movement. However, when you look at historical charts, they have the appearance of being randomly generated as there is no context when viewing historical prices.
    
   
3. Unless we assign useful context to market movements, any analysis which is done is never going to be much better than random.
    
   
4. Would you trade the product in my chart if you knew it were completely random???
   

 

What do you guys reckon?

 

I think that there are both influences of random and non-random price behavior. If the economy goes bad, and earnings reports are bad, then sooner or later, many stock prices will go down. That is not random. It's very basic valuation.

 

But trading and investing has a large degree of speculation to it. And there are many buyers and many sellers. And even though human behavior may be fairly predictable at times, there is a large degree of uncertainty about how people will behave in the short term.

 

I probably would not trade at all, if there was truly total randomness. Although the examples you gave, seem to actually have some structure to them. So I question how random the data generation really is.

[MightyMouse 12]

I think that there are both influences of random and non-random price behavior. If the economy goes bad, and earnings reports are bad, then sooner or later, many stock prices will go down. That is not random. It's very basic valuation.

 

But trading and investing has a large degree of speculation to it. And there are many buyers and many sellers. And even though human behavior may be fairly predictable at times, there is a large degree of uncertainty about how people will behave in the short term.

 

I probably would not trade at all, if there was truly total randomness. Although the examples you gave, seem to actually have some structure to them. So I question how random the data generation really is.

 

People behave exactly the same way every time. They sell when they think price is as high as it is going to go and they buy when they feel that the price is as low as possible. There is nothing random with the above.

[Tradewinds 12]

People behave exactly the same way every time.

 

No, you're wrong. I'm a person, and I don't behave exactly the same way every time. I may be in exactly the same situation twice, and behave differently each time. I might be really, really stupid for doing that. But it's a fact. So you are so wrong! I can't believe how wrong you are! Hopefully you won't be consistently wrong every-time you make a post.

 

Actually, now that I think about it, you might be right. People do behave the same way every time. They are like robots. People think that they are in control of themselves, but they are not. We are all just robots. Complicated robots, but still robots.

 

Some people sell when they want price to be as high as it's going to go. Then they don't get what they want, or they get what they want. Some people sell when their strategy tells them to sell.

 

I know one thing for sure. People are consistently random all the time, in a non-random way. In other words the same patterns of behavior repeat themselves all the time with incredible consistency and predictability within the bigger picture, but trying to guess second to second what the behavior is going to be is more random.

[BlowFish 10]

Okay so here are three(non-cherry picked) 20 bar samples of a 144 tick chart of the random data.

 

Just remember this is a very simple exercise. Each 'tick' which is randomly generated here can give rise to a change in price which is clearly not going to be as likely in a real product. But it does give a nice illustration.

 

A simple price chart does not display all the readily available information about a financial data series. There are three main variables price, time (from the start of the sample or from the last tick) and volume. When you construct a chart you tend to hold one of those constant (to construct a bar). If you display all (e.g a simple time based chart with volume histogram) I think it will be obvious what is random and what is real.

 

Many years ago there was a website you go to and take a test. It simple presented a series of charts (price only) and asked you to select whether it was real or random. (A cursory look and I could not find it today). I remember when I took it there where only a couple of charts that threw me. Real price charts have 'characteristics' that seem pretty rare in random data.

[TheNegotiator 13]

Exactly the sort of thing I am really trying to get at- how should the idea be viewed and is there anything which proves/disproves it and therefore is essential for us to consider? Or does it not matter at all??

 

Anyway, I totally agree about the idea of volume adding a level of context. Remember that a randomiser is just adding price/volume data based on possibilities. Not like traders who are specifically looking at areas of high previous 'value' as a lead to which they may trade off in the future.

 

That being said, I did a quick report in excel with the random data and added random volume at every tick between 1 and 100 and then summed volume at price. I know this isn't necessarily accurate to any particular product but that isn't the point.

 

Just take a look.

[TheNegotiator 13]

Just to be clear on the picture, the volume distribution and the price chart are not the same scale. I have tried to highlight the areas though which the volume distribution covers.

[MightyMouse 12]

No, you're wrong. I'm a person, and I don't behave exactly the same way every time. I may be in exactly the same situation twice, and behave differently each time. I might be really, really stupid for doing that. But it's a fact. So you are so wrong! I can't believe how wrong you are! Hopefully you won't be consistently wrong every-time you make a post.

 

Actually, now that I think about it, you might be right. People do behave the same way every time. They are like robots. People think that they are in control of themselves, but they are not. We are all just robots. Complicated robots, but still robots.

 

Some people sell when they want price to be as high as it's going to go. Then they don't get what they want, or they get what they want. Some people sell when their strategy tells them to sell.

 

I know one thing for sure. People are consistently random all the time, in a non-random way. In other words the same patterns of behavior repeat themselves all the time with incredible consistency and predictability within the bigger picture, but trying to guess second to second what the behavior is going to be is more random.

 

You missed the point. You do not make a decision to enter or exit based on a random decision. It is true that there are many reasons why someone enters or exits bu those reasons are not random. They may appear random to someone who cannot discern the rhythm from the noise. But blindness is no excuse for ignorance.

[TheNegotiator 13]

I did a per bar volume study in the attachment. Although the volume doesn't vary that much, there are 3 of things which should be taken into account explaining this.

 

1- I only randomised the tick volume between 1-100.

 

2- The randomisation is not weighted at all. Most ticks in say ES are small(below 10) so when a couple of bigger clips come in in will change the volume seen in a small tick chart(144 in this case) considerably.

 

3- The bar sample is small given it is only 23 in length.

 

Anyway, here it is.

[tommaso 10]

When I first came across this ages ago it made me sit up and it did the same when I came across it again recently. Take a look at the chart I have attached. What is it you ask? Oil, gold, S&P 500. Nope, it's nothing. It's a chart which I had excel generate simply by getting it to choose a tick higher, a tick lower or unchanged from the last price. But it looks very familiar doesn't it?

 

Now we all know the markets aren't random , but really when you look at an example like this it makes you wonder. What are the implications of such an exercise?

 

Here are some thoughts:

 

1. There is an element of perceived randomness to the market at times which may or may not perpetuate further 'random' activity.
    
   
2. Markets are not random at all as conditions present at any given moment, caused the subsequent price movement. However, when you look at historical charts, they have the appearance of being randomly generated as there is no context when viewing historical prices.
    
   
3. Unless we assign useful context to market movements, any analysis which is done is never going to be much better than random.
    
   
4. Would you trade the product in my chart if you knew it were completely random???
   

 

What do you guys reckon?

 

 

You should first define the notion of Randomness.

 

One of the most famous probabilist, De Finetti, started his famous book on probability with the sentence: "probability does not exist".

 

So, go figure.

 

As we know nothing about future prices, and they are obviously unpredictable, we could say prices are random. Or better, not deterministically predictable. And, actually not even probabilistically. ;-)

 

T

[Jleon 10]

When I first came across this ages ago it made me sit up and it did the same when I came across it again recently. Take a look at the chart I have attached. What is it you ask? Oil, gold, S&P 500. Nope, it's nothing. It's a chart which I had excel generate simply by getting it to choose a tick higher, a tick lower or unchanged from the last price. But it looks very familiar doesn't it?

 

Now we all know the markets aren't random , but really when you look at an example like this it makes you wonder. What are the implications of such an exercise?

 

Here are some thoughts:

 

1. There is an element of perceived randomness to the market at times which may or may not perpetuate further 'random' activity.
    
   
2. Markets are not random at all as conditions present at any given moment, caused the subsequent price movement. However, when you look at historical charts, they have the appearance of being randomly generated as there is no context when viewing historical prices.
    
   
3. Unless we assign useful context to market movements, any analysis which is done is never going to be much better than random.
    
   
4. Would you trade the product in my chart if you knew it were completely random???
   

 

What do you guys reckon?

 

Sure, I would trade it without a doubt, as long as it showed correct price action around my lines Happy trades!

[TheNegotiator 13]

Randomness is definitely about perception in my eyes. The probabilities involved in trading are more about how likely you are to be right and not how likely the market might do something. There are just too many parameters which can't be known in trading at any given time that it's essential to plan for what you don't know about, happening. This is my idea of randomness. It's perception and lack of perception.

[piotrek.gregor 10]

The fact that random generated processes do not fit exactly to what market move is, is well known, although we still use brownian motion etc in quantitative finance.

But I thought that candlestick chart of these random data would be closer to real patterns that we may observe in the market. I think I could tell you, just taking a look at these charts, that these candles are not real data, as was said before: there is a lack of trading patterns in it.

[TheNegotiator 13]

I disagree. I think they do show similarities. Remember the candlestick charts are really too short in length as to be really a good test. Plus I only did a 144 tick chart and I didn't factor in any probability of whether a tick changed the bid/ask or for that matter that the next tick could only be a bid/ask price and not 1 higher/1 lower/unchanged. It's a crude example of randomly generated data which I attempted to make simple but fairly decent, so I don't think picking it apart is really the object of the exercise I had in mind. If anyone with better excel/access skills than me wants to work on a better model for the random data and present it here, that would be great.

[russellhq 10]

Guys. here is an example of random data over 200 days (144 ticks a day).

 

I've attached the spreadsheet used to create the chart. Press F9 and you can refresh the chart to get a new random chart.

RAND.xlsx

[TheNegotiator 13]

I don't know exactly how you generated the data, but if it was well thought through, the chart you show is a good demonstration. It shows consolidation followed by a break then a retest of consolidation followed by a trend then a double top!! No patterns huh?

[TheNegotiator 13]

Looking at it again, I was actually quite suspicious as it looks so real. I looked at the file though and generated some more charts. It's a really good illustration.